Scott please try following the order sudo ethtool -A dna0 autoneg off sudo ethtool -A dna0 rx off sudo ethtool -A dna0 tx off
Alfredo On Nov 30, 2011, at 2:28 AM, Scott L wrote: > Alfredo, > > By simultaneously watching the switch's statistics and pfcount, we see > that initially the switch sees no pause frames and pfcount sees > packets. When pfcount stops receiving packets, the switch immediately > starts receiving lots and lots of pause frames. At this point, the > only known remedy is to reload the kernel modules using > load_dna_driver.sh. > > Turning autonegotiation off doesn't change anything. I can't seem to > turn off anything else (see below) > > [examples]$ sudo ethtool -A dna0 rx off > [examples]$ sudo ethtool -a dna0 > Pause parameters for dna0: > Autonegotiate: on > RX: on > TX: on > > [examples]$ sudo ethtool -A dna0 tx off > [examples]$ sudo ethtool -a dna0 > Pause parameters for dna0: > Autonegotiate: on > RX: on > TX: on > > =s= > > On Tue, Nov 29, 2011 at 6:07 PM, Alfredo Cardigliano > <[email protected]> wrote: >> Scott >> it is possible that the flow control slows down a link, but in your case it >> just stops. >> If you want you can try with these commands to see if something changes: >> ethtool -A dna0 autoneg off >> ethtool -A dna0 off >> ethtool -A dna0 tx off >> >> Alfredo >> >> On Nov 30, 2011, at 1:07 AM, Scott L wrote: >> >>> Alfredo, >>> >>> We looked at the statistics produced by our switch and discovered that >>> when pfcount stops reporting received packets, we start seeing Pause >>> frames from the Silicom NIC. This would account for the fact that we >>> stop seeing packets. Have you seen this issue before? Any idea how >>> we might resolve it? >>> >>> =s= >>> >>> On Tue, Nov 29, 2011 at 10:46 AM, Alfredo Cardigliano >>> <[email protected]> wrote: >>>> Scott >>>> actually we recommend kernels > 2.6.18 with vanilla PF_RING, not with DNA. >>>> However 2.6.31 should be fine. >>>> Upgrading to a newer version is not a bad idea, if possible. Currently we >>>> are mainly using 2.6.36 and newer kernels. >>>> A asked you about 32/64bit just because there was an old 32bit dna library >>>> in svn. >>>> >>>> Regards >>>> Alfredo >>>> >>> _______________________________________________ >>> Ntop-misc mailing list >>> [email protected] >>> http://listgateway.unipi.it/mailman/listinfo/ntop-misc >> >> _______________________________________________ >> Ntop-misc mailing list >> [email protected] >> http://listgateway.unipi.it/mailman/listinfo/ntop-misc > _______________________________________________ > Ntop-misc mailing list > [email protected] > http://listgateway.unipi.it/mailman/listinfo/ntop-misc _______________________________________________ Ntop-misc mailing list [email protected] http://listgateway.unipi.it/mailman/listinfo/ntop-misc
